Reviews, Editor Tests or Cosmetic Claims Evidence: Which Beauty Evidence Can You Actually Trust?

Everything in beauty is presented as evidence. A review with five stars, a magazine's tested and approved badge, a dermatologist on social media, a percentage on a box, a before and after photograph, a founder's story. Some of it is evidence, some is experience and some is decoration, and the difference is not obvious from the packaging or the feed. Underneath all of it sit two documents that the law requires to exist and to be produced on request: the Cosmetic Product Safety Report required by Regulation (EC) No 1223/2009 and enforced in Great Britain through the Cosmetic Products Enforcement Regulations 2013, and the Product Information File holding evidence for every claim under Regulation (EU) No 655/2013. Most consumers have never heard of either.

This article ranks the sources. It explains what the two regulated documents actually prove and who can demand them, then works down through instrumental studies, perception studies, editor tests, consumer reviews, before and after photographs, influencer content and founder narratives, saying what each can and cannot establish. It covers the specific failure modes of each source, the three questions that fact check almost any beauty claim in a minute, how the regulators handle complaints and how quickly, a worked example of one product's evidence assessed from the outside, and what brands should publish if they want to be trusted by people who have learned to check.

The hierarchy, top to bottom

What the CPSR actually proves, and what it does not

It proves safety, quantitatively. For each ingredient with a systemic endpoint the assessor calculates exposure from the product and compares it with the level at which no adverse effect was observed, normally requiring a margin of at least 100, and the report concludes only if the product is safe under its conditions of use. It also records whether the claims made for the product are consistent with the assessment, which is how medicinal claims and impossible claims get struck before printing. This is also where the six common criteria for cosmetic claims come in, since an efficacy claim that fails them, or that implies a physiological function rather than a cosmetic effect, risks tipping the product into medicinal product territory. What it does not do is prove that the product works. Efficacy evidence is separate, held in the Product Information File, and a brand with a beautiful Cosmetic Product Safety Report may have no performance evidence at all, which is entirely lawful provided it makes no unsupported performance claims, such as a claim to reduce the appearance of wrinkles that no study actually measured. Understanding that division is the single most useful thing a consumer can learn about beauty documentation, and our article on clean beauty and the CPSR goes further into the safety side.

Why reviews are weak evidence for effect

Not because reviewers are dishonest, though some are paid and some are fabricated, but because a review cannot isolate the product. A person starting a new serum has usually also changed something else: the season, their sleep, their water intake, another product, or simply their attention to their skin. Expectation shapes perception strongly in cosmetics, which is why perception data without controls are treated cautiously. Timing distorts too, since most reviews are written in the first fortnight, before effects that take eight weeks could appear and after the immediate sensory experience has peaked. And the distribution is unreliable: delighted and furious customers write, while the majority who found the product adequate do not. Reviews are genuinely useful for things they can observe directly, such as whether a pump works, how a texture feels, whether the scent is strong and whether delivery was quick. For a measured biological effect they are the wrong instrument, and a review is never solid evidence for a clinically proven or dermatologically tested claim in the way a proper controlled study would be.

Editor tests and awards

These vary enormously and should be read for what they are. At their best, a publication tests a category on a panel over a period, describes the method, discloses commercial relationships and publishes results including products that did not perform. At their weakest, an award is a paid entry scheme with a fee for using the logo, or a roundup written from press releases. Neither extreme is dishonest if disclosed, but only the first is evidence, and even then it is evidence of a panel's impressions rather than a measured effect. The useful question is whether the publication describes what it did. A tested on 20 readers over four weeks with photographs assessed by a dermatologist is a real method; a best cleanser 2026 badge with no method is a recommendation, which may still be a good one but is not a finding, and treating it as scientific evidence of efficacy would mislead consumers. Disclosure rules apply where an award or a review is advertising, and the Advertising Standards Authority has ruled on undisclosed arrangements and on misleading claims more generally, including those presented as independent editorial judgement when they were paid placements.

Before and after photographs

The most persuasive and least reliable format in beauty. Lighting direction and colour temperature, camera angle, distance, focal length, expression, hydration on the day, makeup, editing and simply which of forty shots was chosen can produce a dramatic difference with no change to the skin at all. Where before and after images are used in advertising, the Advertising Standards Authority expects them to be representative and not exaggerate the effect, and requires that any post production be disclosed or avoided. For a consumer the practical filter is consistency: are the lighting, angle, expression and framing identical, is the timescale stated, and is the image from a study with a panel rather than a single customer. If any of those is missing, treat the pair as illustration rather than evidence.

Influencer content is advertising

When content is paid for, gifted, commissioned, or created under an affiliate arrangement, it is advertising, and the same claims rules apply to it as to the brand's own packaging. That means a medicinal claim in an influencer video is a medicinal claim by the brand, a percentage quoted without a study is an unsupported claim, and disclosure must be clear and upfront rather than buried in hashtags. Brands often discover this when a complaint is made about a post they did not write, and the answer is a written brief with approved and prohibited wording, which good agencies expect. For consumers, the useful habit is to look for the disclosure first and the specifics second: a creator describing texture and finish is giving you their experience, and a creator quoting a percentage should be asked where the number came from, because usually the brand supplied it and sometimes the study does not exist.

The three question fact check

• Proven what, exactly? A specific endpoint, not a vibe. Reduced appearance of wrinkles is a claim; transforms your skin is not.
• Measured how? An instrument, a photograph assessed by an expert, or a questionnaire. Each supports different wording.
• On how many people, for how long? A percentage from twenty participants over two weeks is a very different number from one from eighty over eight.

A brand that can answer all three in one email has evidence. A brand that answers with adjectives has marketing. This works on any claim, from any source, and takes about a minute.

How do complaints actually work?

Faster than most brands expect. The Advertising Standards Authority accepts complaints from anyone, including competitors, and asks the advertiser for substantiation within a short deadline, typically days rather than weeks; rulings are published and searchable. The Competition and Markets Authority handles environmental claims under the Green Claims Code and has acted on fake and incentivised reviews. Trading Standards and OPSS can require the Product Information File, where the claims evidence must already be held. And the MHRA acts where a cosmetic is presented as a medicine. The common thread is that the evidence has to exist at the time the claim is made; there is no period in which to go and commission a study after the complaint arrives, which is why the claims table matters more than any individual piece of marketing.

Worked example: one product assessed from the outside

A serum claims clinically proven to reduce fine lines by 42 percent, dermatologist approved, 96 percent of users agreed, with two before and after images and 4.8 stars from 3,000 reviews. Applying the hierarchy takes ten minutes. The 42 percent is the strongest of the efficacy claims and the brand's page states it came from an eight week instrumental study on 44 volunteers with image analysis, which is real evidence and unusually well disclosed. Dermatologist approved has no method described, so an email is sent asking what was assessed, since worded carelessly it could edge toward sounding like a medical claim rather than a cosmetic one. The 96 percent turns out, on the same page, to be from 25 participants, which the brand states, so it is honest but modest. The images have different lighting and angle and are treated as illustration. The reviews are useful for texture and the pump, and irrelevant to the wrinkle claim. Conclusion: one strong claim with a study, one honest small perception claim, one unverified endorsement and a great deal of decoration. That is a better than average result, and the brand's willingness to publish its panel sizes is the reason it can be checked at all.

What brands should publish to be trusted

• The full ingredient list on every product page. 
• What each key ingredient does and at what level. 
• For every performance claim, the endpoint, method, duration and panel size. 
  Whether a study was on the finished product or an ingredient. 
• Panel size alongside every percentage. 
• Clear disclosure on gifted and paid content, with a written brief behind it. 
  Who assessed the safety of the product, and that a Product Information File exists. 
•  A contact that answers specific evidence questions within a working day. 
•  A plain statement of what the product does not do.Brands that say a serum will not replace sun protection, or that a cosmetic cannot regrow hair, are the ones customers believe on everything else.

Myths we hear

• Thousands of reviews prove a product works. They evidence experience and popularity, not effect.
• A magazine award is independent verification. Sometimes; ask what the method was and whether entry was paid.
• Influencer posts are the creator's opinion, not our claim. When gifted or paid, they are advertising and the brand is responsible.
• Before and afters speak for themselves. They speak for lighting and angle unless controlled.
• We can substantiate if challenged. The evidence must exist when the claim is made, and deadlines are days.
  

What to do this week

• Apply the three question test to the strongest claim in your own marketing. 
• Add panel sizes to every percentage you publish. 
• Write an influencer brief with approved and prohibited wording. 
• Check that your before and after images are consistent and disclosed. 
• Make sure the claims table in your Product Information File is current. 

Beauty does not lack evidence; it lacks the habit of showing it. The brands that publish their methods are the ones that survive being checked, and being checked is becoming normal.

Related reading on the Spectra blog

•    Clinically proven and dermatologist tested: the evidence a CPSR files
•    How a Cosmetic Product Safety Report exposes greenwashing
•    Clean beauty and the Cosmetic Product Safety Report
•    All Spectra articles

Spectra services mentioned in this article

•    The Complete Compliance Pack: CPSR, PIF, SDS and SCPN or CPNP (£295)
•    CPSR services by product type
•    PIF services
•    UK and EU Responsible Person
•    Contact Spectra

Official sources and legislation

•    Regulation (EU) No 655/2013 on cosmetic claims
•    Advertising Standards Authority
•    CMA Green Claims Code
•    Regulation (EC) No 1223/2009 on cosmetic products
•    OPSS guidance on cosmetic products regulations

Disclaimer: This article is general information, not legal advice. Regulations change and product decisions depend on your exact formulation, claims and market. Have your product assessed by a qualified professional before sale.